"That's right, despite the fact the republicans haven't had control over the government for nearly 10 years"
I do not follow American politics that much, but I was under the impression that Bush was republican and still in office in 2008. Did I miss something?
I, hereby, declare january 29th "birthday of somersault's sister day"! At least we will have a "something-day" one person gives shit about.
Re:What does the hell does NP Hard mean?
on
Pac-Man Is NP-Hard
·
· Score: 1
I read so much wrong answer replying to you post I attempt a new explanation.
A problem is in P if there is a polynomial time algorithm to solve it. It means, that there is a algo which will always find the correct answer AND the algorithm an instance of the problem in at most a*n^b operations, where a and b are constants and n is the size of the smallest file that can contain an instance of the problem. Example of problems in P includes: deciding whether a number is divisible by2, whether a number is prime, finding a shortest path between two points in a city, solving linear equations,...
A problem is in NP, if a solution of an instance of a this problem has a solution whose size is at most a*n^b. I do not care about algorithms that generate that solution, juste about the size of the solution. Interestingly, all problem that are in P are also in NP. The reverse, might or might not be true.
A problem is NP-hard if, assuming we have an algorithm that can solve that problem in constant time for (think one assembly instruction on your computer), it can be used to solve all the problems in NP in polynomial time.
If a problem is NP-hard AND in NP, then the problem is said to be NP-complete. This class of problem is interesting because, some mathematical techniques have been used to show that some practically relevant problems are NP-complete: the most comon one if the post delivery problem: "is it possible to deliver the mail for all the inhabitant of a given city while travelling less than a given number of miles".
There are no known polynomial algorithms for NP-complete problems. And actually, most theoreticians believe that we do not know any because no such algorithm exist. So if you have an algorithm that solve all problems in a*n^b operations, then b is typically not a constant, it is a function of n. (or at least we believe)
Notice that all the talk on P, NP and NP-complete consider ALL the instances of a given problem. Some instances might be easy for some algorithms but if a problem is NP-complete (or NP hard), then there exists an instance which the algorithm can not solve in polynomial time (or at least that what we believe).
Interestingly, all NP-complete problems are "equivalent", if we can solve one in polynomial time, we can solve them all in polynomial time. That mandate the search for weird problems that can be shown to be NP-complete. Because solving any of them in polynomial time would solve them all. And we have been trying to do that for more than 60 years.
Moreover, because of how NP is defined, all NP-complete problems can be solved by an exponential algorithm (one that would take less than a*n^(b*n^c), where a, b and c are constants)
P-space is a little bit different. All the classes I mentionned before, P, NP, NP-Complete are interested in runtime of the algorithm and are sometimes called P-time, NP-time, NP-time-complete. P-space is interested in how much memory is required to solve a problem. Interestingly, P-time and NP-time are in P-space.
Showing that doom is P-space-hard is interesting because there are not too many practically relevant problems that are shown to be P-space-hard.
Trying to close the analog hole I guess. Using "smart" HDMI can more easily be used with DRMs. Coupled with machine you can not choose the OS of, and you might have quite annoying copy protection schemes.
I guess the war has now begun. Taking down the department of justice is a clear start of all hostility. I am not sure I agree with them. But they have stuff in their pants!
Personally, i wish the next GPL license would be more open, in the sense of: you can protect this very piece of code, but remove the 'viral' part from it, and allow usage in any further license. Example are the BSD unices that complain how they cannot use GPL'd drivers and other code, whereas the 'linux/gnu team' can happily borrow code from them. That's up to the very GPL license to fix. Same applies to closed source - i bet a company would happily show code of a GPL library they used, and possibly modified, if it was only limited to that.
What you describe is pretty much the same thing as LGPL or the license of the C++ standard lib that comes with gcc. It is a GPL license with exception clause to avoid GPL contamination because of template and header inclusion.
Which is exactly why I love the GPL. I know the code I will release might be reused by other people but they MUST release their modification of it. That way my effort enters a public pool of knowledge and so will every derivate of what I do. I release my code under GPL to expand the public knowledge and code base. I know all modifications will also expand public knowledge.
I am tired of business monopolies. Phone carriers have monopolies because the billion dollars infrastructure needed to even start being competitive. Mobile phone providers have monopolies because of million of patent you need to license before doing anything remotely close to a phone. Same goes for pretty much all computers and objects more complex than a table. Software is a place where we can fight monopolies easily by releasing GPL code. It provides prior art to many patents and decrease significantly the cost of entering in business. They provide reference/learning implementation for pretty much anything. Entering business becomes easy and reasonnably cheap.
Long live the GPL and fuck businesses and their so called "trade secret"
Some more anecdotal evidence here. I know 4 people with xboxes using them for times between 1 year and 4 years. None had hardware problems. So I am not too sure. Are there actual statistics on that issue?
marking the bug closed because it can not be reproduced is a good way of carrying the right message. You can also attach reasons why you can not reproduce it by listing missing details. With time the users might learn to make a bug report. Of course there should be an option to allow the user to reopen the bug.
If they are smart they can make WebOS, the mobile phone platform that RMS can approve of. And then they'd have something of value.
Unfortunately, that does not count for much in the regular world... Even in the geek world, the amount of people that would switch from android to something else just based on license would be minimal. I am pretty opiniated on licensing issues, but moving from android (a hacked version) to webOS does not make much sense...
no, the genome of every bacteria in a soil sample. We (I work as a computer scientist in a genome related research lab) do not work only on human genomes.
I installed the server on my machine and gave it a shot. I made very classical request such as the name of a couple universities, a couple famous website and made a few regular queries like "chocolate mousse recipe". None of the request actually pointed to something even remotely close to what I was looking for. I thought it might need some bootup time, so I tried again an hour later. It was not much better. Just much slower. I'll try again in a few day. But that does not look good...
On top of that it looks like there is no special ranking system, so I guess they take the order of reply or the number of occurences... most likely not good criteria...
"I still have to ask the general public whether, desktop Linux still matters. Does it?"
The general public has no clue linux actually exists. But there remain a part of the population (0.1%) that never use anything else than linux. I do not recall when was the last time I used a windows machine for more than an hour. I think it was somewhere in 2006.
Most likely that part of the population read slashdot:)
There is a difference between "socially valuable" and "in demand". I think it is a good thing to have many history major or art major. I love talking with them and I think they contribute positively to society. But somehow they are not "in demand".
By having education pushed by loans, you do not remove the segregation problem. Many people drop out of college or somewhere in their graduate study because they need the money to pay for their loan. They end up needing to pay back unrealistically high amount of money. If your parents have money, the problem is not so big. The segregation is still there.
I believe having universities be publicly funded allows to level the field of who can reach higher level of education based on their natural abilities and desire. Public fund are not unlimited as well, you can then filter student based on abilities and motivations.
I have been told there is a public grant funding opportunities in the USA. But I never really saw it work properly. (Note that I haven't looked to much into it as well.)
I usually need a lot of tools because I have a versatile job. As a researcher in a university in a close R&D department, I often have to test tools and analyse data that come a little bit from everywhere.
Often I have root access on my machines. Once I did not have root privilege on my desktop because of "security policy". I ended up asking IT to install software frequently. For some reason the IT guy believed he could do my job better than me and knew which tools I need better than me. Every time, the IT asked me stupid question, like "why do you need an installation of pdflatex? you have latex already!!" "well, the journal we are submitting to uses pdflatex and our article does not compile." "In my experience, journal use latex" "!? well, this one doesn't" "I see. Why don't you install it on your home directory?" "I could, but installing a latex distribution manually is a nightmare. As root, it only requires installing one package and let the package manager do its job. In 10 minutes it is installed, properly configured and will update automatically with the system." "Latex is not updated very often, so the automatic updates are not very useful. You could install it based on a chroot in your home directory"... it went like that for about 20 minutes Two days later: "Could you install ruby on our computing nodes?" "Why do you need ruby? It is not a very good programming language and it is significantly less efficient than alternatives like python." "Because I need tool-foo which is written in ruby." "Oh I see. Instead of tool-foo, you could use tool-bar which is written in java and does almost the same thing." "Well, I need tool-foo because tool-bar does not have a feature I need." "Which feature? In my experience users ask for many different tools without wondering if another tool happen to have the proper features."... It went like that for 30 minutes. the week after "Could you install git on my machine and on our computing nodes?" "Why do you need git?" "To have versioning of my code and experiement" "We have an svn server, why don't you use that?" "because the svn server has a limited capacity and it relies on accessing the network, which is not accessible on our computing nodes. But git is point to point and works great over ssh." "I see. I guess we could set up a git server to synchronize the machine..." "Well, I don't need a git server. I just need the git package to be installed" "... so I need to install a new virtual machine. But I will need to connect it to the LDAP. Oh yes the problem of accessing from the computing nodes, so I could modify the settings of the firewall..." "I don't need a git server I just need git. I'll synchronize on the file system" "... but if I change the setting of the firewall, you could access the SVN server. So why don't you use SVN?" "because the SVN server will never support the load I am going to push to the repository" "I see. In my experience, people in university use git mainly to contribute to open source software and not for actually working." "... *sigh*"
I let you imagine the day I requested a kernel update...
Because otherwise education is only available to rich people. It means you filter people to get a high education based on the money their parent have instead of the natural ability of the kid. It is first extremely unfair (but you would classify that as a socialist problem) but it also mean that you prevent very smart people to get a reasonnable education and contribute positively to the society. Instead they are going to work for walmart.
It depends what I read and why I read it. If I want some high level factual description it is pretty much what I do. If I want to really read a book and immerse myself in the book, I read around 1 page per minute to.7 page per minute. when reading something technical, reading is not the bottleneck.
Slashdot Mirror
http://www.enterprisestorageforum.com/print/storage-hardware/linux-file-system-fsck-testing----the-results-are-in.html
going through 3 pages is so annoying...
"That's right, despite the fact the republicans haven't had control over the government for nearly 10 years"
I do not follow American politics that much, but I was under the impression that Bush was republican and still in office in 2008. Did I miss something?
"either leave food and especially bathrooms out of the comparison or exclude the UK from Europe."
Here is a good insight into British food!
I, hereby, declare january 29th "birthday of somersault's sister day"!
At least we will have a "something-day" one person gives shit about.
I read so much wrong answer replying to you post I attempt a new explanation.
A problem is in P if there is a polynomial time algorithm to solve it. It means, that there is a algo which will always find the correct answer AND the algorithm an instance of the problem in at most a*n^b operations, where a and b are constants and n is the size of the smallest file that can contain an instance of the problem. Example of problems in P includes: deciding whether a number is divisible by2, whether a number is prime, finding a shortest path between two points in a city, solving linear equations,...
A problem is in NP, if a solution of an instance of a this problem has a solution whose size is at most a*n^b. I do not care about algorithms that generate that solution, juste about the size of the solution. Interestingly, all problem that are in P are also in NP. The reverse, might or might not be true.
A problem is NP-hard if, assuming we have an algorithm that can solve that problem in constant time for (think one assembly instruction on your computer), it can be used to solve all the problems in NP in polynomial time.
If a problem is NP-hard AND in NP, then the problem is said to be NP-complete. This class of problem is interesting because, some mathematical techniques have been used to show that some practically relevant problems are NP-complete: the most comon one if the post delivery problem: "is it possible to deliver the mail for all the inhabitant of a given city while travelling less than a given number of miles".
There are no known polynomial algorithms for NP-complete problems. And actually, most theoreticians believe that we do not know any because no such algorithm exist. So if you have an algorithm that solve all problems in a*n^b operations, then b is typically not a constant, it is a function of n. (or at least we believe)
Notice that all the talk on P, NP and NP-complete consider ALL the instances of a given problem. Some instances might be easy for some algorithms but if a problem is NP-complete (or NP hard), then there exists an instance which the algorithm can not solve in polynomial time (or at least that what we believe).
Interestingly, all NP-complete problems are "equivalent", if we can solve one in polynomial time, we can solve them all in polynomial time. That mandate the search for weird problems that can be shown to be NP-complete. Because solving any of them in polynomial time would solve them all. And we have been trying to do that for more than 60 years.
Moreover, because of how NP is defined, all NP-complete problems can be solved by an exponential algorithm (one that would take less than a*n^(b*n^c), where a, b and c are constants)
P-space is a little bit different. All the classes I mentionned before, P, NP, NP-Complete are interested in runtime of the algorithm and are sometimes called P-time, NP-time, NP-time-complete. P-space is interested in how much memory is required to solve a problem. Interestingly, P-time and NP-time are in P-space.
Showing that doom is P-space-hard is interesting because there are not too many practically relevant problems that are shown to be P-space-hard.
I hope it helps people understand.
NP-hard Pacman's got nothing on Undecidable Ms. Pacman.
There is a known theorem that says: "As soon as you throw women in, nothing is decidable anymore!"
Trying to close the analog hole I guess. Using "smart" HDMI can more easily be used with DRMs. Coupled with machine you can not choose the OS of, and you might have quite annoying copy protection schemes.
I guess the war has now begun. Taking down the department of justice is a clear start of all hostility. I am not sure I agree with them. But they have stuff in their pants!
Isn't the raspberry pi significantly more powerful and cheaper?
Personally, i wish the next GPL license would be more open, in the sense of: you can protect this very piece of code, but remove the 'viral' part from it, and allow usage in any further license. Example are the BSD unices that complain how they cannot use GPL'd drivers and other code, whereas the 'linux/gnu team' can happily borrow code from them. That's up to the very GPL license to fix. Same applies to closed source - i bet a company would happily show code of a GPL library they used, and possibly modified, if it was only limited to that.
What you describe is pretty much the same thing as LGPL or the license of the C++ standard lib that comes with gcc. It is a GPL license with exception clause to avoid GPL contamination because of template and header inclusion.
Which is exactly why I love the GPL. I know the code I will release might be reused by other people but they MUST release their modification of it. That way my effort enters a public pool of knowledge and so will every derivate of what I do. I release my code under GPL to expand the public knowledge and code base. I know all modifications will also expand public knowledge.
I am tired of business monopolies. Phone carriers have monopolies because the billion dollars infrastructure needed to even start being competitive. Mobile phone providers have monopolies because of million of patent you need to license before doing anything remotely close to a phone. Same goes for pretty much all computers and objects more complex than a table. Software is a place where we can fight monopolies easily by releasing GPL code. It provides prior art to many patents and decrease significantly the cost of entering in business. They provide reference/learning implementation for pretty much anything. Entering business becomes easy and reasonnably cheap.
Long live the GPL and fuck businesses and their so called "trade secret"
Some more anecdotal evidence here. I know 4 people with xboxes using them for times between 1 year and 4 years. None had hardware problems. So I am not too sure. Are there actual statistics on that issue?
marking the bug closed because it can not be reproduced is a good way of carrying the right message. You can also attach reasons why you can not reproduce it by listing missing details. With time the users might learn to make a bug report. Of course there should be an option to allow the user to reopen the bug.
If they are smart they can make WebOS, the mobile phone platform that RMS can approve of. And then they'd have something of value.
Unfortunately, that does not count for much in the regular world... Even in the geek world, the amount of people that would switch from android to something else just based on license would be minimal. I am pretty opiniated on licensing issues, but moving from android (a hacked version) to webOS does not make much sense...
I want to ask slashdot what kind of clothes I should wear.
Sorry, there is no expert on slashdot for these type of questions!
no, the genome of every bacteria in a soil sample. We (I work as a computer scientist in a genome related research lab) do not work only on human genomes.
I installed the server on my machine and gave it a shot. I made very classical request such as the name of a couple universities, a couple famous website and made a few regular queries like "chocolate mousse recipe". None of the request actually pointed to something even remotely close to what I was looking for. I thought it might need some bootup time, so I tried again an hour later. It was not much better. Just much slower. I'll try again in a few day. But that does not look good...
On top of that it looks like there is no special ranking system, so I guess they take the order of reply or the number of occurences... most likely not good criteria...
"I still have to ask the general public whether, desktop Linux still matters. Does it?"
The general public has no clue linux actually exists. But there remain a part of the population (0.1%) that never use anything else than linux. I do not recall when was the last time I used a windows machine for more than an hour. I think it was somewhere in 2006.
Most likely that part of the population read slashdot :)
two comments on that:
There is a difference between "socially valuable" and "in demand". I think it is a good thing to have many history major or art major. I love talking with them and I think they contribute positively to society. But somehow they are not "in demand".
By having education pushed by loans, you do not remove the segregation problem. Many people drop out of college or somewhere in their graduate study because they need the money to pay for their loan. They end up needing to pay back unrealistically high amount of money. If your parents have money, the problem is not so big. The segregation is still there.
I believe having universities be publicly funded allows to level the field of who can reach higher level of education based on their natural abilities and desire. Public fund are not unlimited as well, you can then filter student based on abilities and motivations.
I have been told there is a public grant funding opportunities in the USA. But I never really saw it work properly. (Note that I haven't looked to much into it as well.)
I usually need a lot of tools because I have a versatile job. As a researcher in a university in a close R&D department, I often have to test tools and analyse data that come a little bit from everywhere.
Often I have root access on my machines. Once I did not have root privilege on my desktop because of "security policy". I ended up asking IT to install software frequently. For some reason the IT guy believed he could do my job better than me and knew which tools I need better than me. Every time, the IT asked me stupid question, like ... it went like that for about 20 minutes ... It went like that for 30 minutes.
"why do you need an installation of pdflatex? you have latex already!!"
"well, the journal we are submitting to uses pdflatex and our article does not compile."
"In my experience, journal use latex"
"!? well, this one doesn't"
"I see. Why don't you install it on your home directory?"
"I could, but installing a latex distribution manually is a nightmare. As root, it only requires installing one package and let the package manager do its job. In 10 minutes it is installed, properly configured and will update automatically with the system."
"Latex is not updated very often, so the automatic updates are not very useful. You could install it based on a chroot in your home directory"
Two days later:
"Could you install ruby on our computing nodes?"
"Why do you need ruby? It is not a very good programming language and it is significantly less efficient than alternatives like python."
"Because I need tool-foo which is written in ruby."
"Oh I see. Instead of tool-foo, you could use tool-bar which is written in java and does almost the same thing."
"Well, I need tool-foo because tool-bar does not have a feature I need."
"Which feature? In my experience users ask for many different tools without wondering if another tool happen to have the proper features."
the week after
"Could you install git on my machine and on our computing nodes?"
"Why do you need git?"
"To have versioning of my code and experiement"
"We have an svn server, why don't you use that?"
"because the svn server has a limited capacity and it relies on accessing the network, which is not accessible on our computing nodes. But git is point to point and works great over ssh."
"I see. I guess we could set up a git server to synchronize the machine..."
"Well, I don't need a git server. I just need the git package to be installed"
"... so I need to install a new virtual machine. But I will need to connect it to the LDAP. Oh yes the problem of accessing from the computing nodes, so I could modify the settings of the firewall..."
"I don't need a git server I just need git. I'll synchronize on the file system"
"... but if I change the setting of the firewall, you could access the SVN server. So why don't you use SVN?"
"because the SVN server will never support the load I am going to push to the repository"
"I see. In my experience, people in university use git mainly to contribute to open source software and not for actually working."
"... *sigh*"
I let you imagine the day I requested a kernel update...
Because otherwise education is only available to rich people. It means you filter people to get a high education based on the money their parent have instead of the natural ability of the kid. It is first extremely unfair (but you would classify that as a socialist problem) but it also mean that you prevent very smart people to get a reasonnable education and contribute positively to the society. Instead they are going to work for walmart.
"What happened to the Pentium 5 through 349?"
Their marketing guy comes from mozilla corporation.
I'll agree with you when the robot will argue the question.
... if I did not had to use an eye-tracking device!
It depends what I read and why I read it. If I want some high level factual description it is pretty much what I do. If I want to really read a book and immerse myself in the book, I read around 1 page per minute to .7 page per minute.
when reading something technical, reading is not the bottleneck.